Statistics and biostatistics

By Dr.Nihal Salah Shihab
Prof. Of Public Health &
Community medicine

Observational Studies
Experimental Studies

Observational studies
Experimental or
intervention studies
The study use existing
phenomena to
understand aspects of
health or illness
Investigators study
people and exposures “in
nature
The study test the effect of
some intervention on a
certain aspect of health or
illness.
Investigator assigns
(intervene) treatment or
exposure by Random allocation of
study subjects

Observational studies Experimental or
intervention studies
The investigator
neither controls the
population nor
factors exposed
The investigator
can control both
the population &
factors exposed

Observational
studies
Descriptive
Look at Distribution of
disease+ no
Comparison group
Analytic
Look at Determinants
of disease + have a
comparison group
Experimental
(intervention) or Quasi-
experimental Studies
Randomized
controlled Trial,
Community trial &
lab.trial

1. Health care planning
2. Hypothesis generation (constitutes an important
first step in the search for determinants or risk factors)
3. Trend Analysis

Descriptive
study
Case report+
case series
Cross
sectional
study
Ecological
Correlational
study
Analytic
study
Case control
study
Cohort
study
Observational
No comparison
Group
Have a comparison
Group

1-Case Report 2-Case series
Report a new or unique
finding in a disease e.g.
Previously un-described
disease
Unexpected link between
diseases
Unexpected new
therapeutic effect
A rare feature in a disease
Describe Unusual
characteristics of a
number of patients
with a given disease

1- Case Report 2-Case series
Example:
Pulmonary
embolism among
premenopausal
female.
Example:
Kaposi sarcoma a
disease of old age
has been reported
in young adult
homosexual men .

1-Case Report 2-Case series
Strengths and
limitation:
It provides the first
report of unexpected
event,
It Generate hypotheses
for testing and for
further study.
Strengths and
limitations:
It describe unusual
variation of a disease
It can generate
hypothesis.

1-Case report 2- Case series
Generalization of
the results are
rarely
Generalization of the
results is
limited because:
1-Selection of study
subjects is
unrepresentative
2-No comparison
group

An ecological study is an epidemiological
study in which the unit of analysis is a
population rather than an individual.
(uses data from the entire population to
compare disease frequencies in relation to
certain exposure during the same period of
time or at different points in time.

It correlates between variables
Example:
the relation between exposure to black cloud and
chest infection in different cities
Lack of water sources flourides and dental cares in
different communities or countries
The association between industrialization and
obstructive Pulmonary diseases or lung cancer in
different countries
comparing several governorates based on state-
wide average air pollution and state-wide average
prevalence of respiratory diseases)

Strengths
Easy
Quick
In-expensive
Data about disease
frequency or
prevalence are available

They cannot prove that
one variable causes a
change in another
variable.
(disease causation)
Does not allow for
scientific control of
variables.
Data of
exposure/health is
Unknown for
each individual

This is the observation of a defined
population at a single point in time or
time interval.
Exposure and outcome are measured at
the same time point.
It Collect data on a health problem from a
group of subjects at a single point of time
.

Study question :
Occurrence of disease in a single
point of time
Example :
Number of female using contraceptive
pills among those attending maternal
and child health center

Advantages Advantages
Serengths
Easy, quick and not
expensive.
Estimate the overall
prevalence of a
disease
Estimate exposure
proportion in the
population
Suitable to study
prevalence of chronic
disease.
The first step to
develop evidence for
causal association.

Limitations
 Not useful for
establishing causal
relationships.
It does not allow us to
answer the question
,which came first
(exposure or the disease)
It may be a coincidence
Data are susceptible
to distortion through
the bias ( ‫وانحراف‬ ‫انحياز‬
in the research during
sampling,
questionnaire and
interviewing.
Limitations

The population census.
Market research organizations often use cross sectional
studies (for example, opinion polls), the type of sample
used here may be a convenience sample or a quota
sample.
The prevalence study of skull fractures in children
admitted to hospital in Edinburgh from 1983 to 1989.
Note that although the study period was seven years it
was not a longitudinal or cohort study because
information about each subject was recorded at a single
point in time.

Descriptive
study
Case report+
case series
Cross
sectional
study
Ecological
Correlational
study
Analytic
study
Case control
study
Cohort
study
Observational
No comparison
Group
Have comparison
Group

What is the study?
It is an observational study that
Provide a simple way to investigate
causes of diseases,
especially rare ones. (behcet syndrome),
Arthritis, Juvenile reumatoid,( cystic fibrosis)

Study question:
Does the present outcome (disease), occurred
due to certain exposure
In this study the investigator starts with the
outcome and look backward to detect the history
of exposure to the suspected cause.

Identify the cases according to standard criteria for
diagnosis (study group) (case definition)
Control subjects must be:
Free from the studied disease and should be
Representative of the source population from which
cases were derived.
They are Matched with cases especially for characteristics
that may affect the disease occurrence e.g. age, sex,
social class…..
Ex, a case control study of asthma among high school
students .

Advantages of case
control study:
Disadvantages of case
control study:
Easy, quick, and cheap.
Useful in studying rare
diseases (breast cancer
in men)
Allows the study of
several risk factors
Can prove that an
exposure is related to an
outcome
Can’t calculate rates
Not useful in rare
exposure.
Recall bias as it
depends on
retrospective history.

A measure of association
Can be used to:
Measure effect size or quantify the strength of
association
between a potential risk or protective factor
(exposure) and an outcome.
Odds ratio (O.R):
O.R. = Odds of exposure among cases / Odds of
exposure among control

The result of dividing the proportion of
individuals with an attribute by the proportion
of individuals without the attribute in a
population .
The probability of occurrence of an event as
compared with (divided by) the probability of
non-occurrence of the event in a population.

TotalControlCases
BAExposed
DCNot exposed
B / DA / C ÷

So, ex, among cases the odds would be
A/A+C ÷C/A+C = A/C
So, ex, among control the odds would be
B /B+D ÷ D/ B+D = B /D
a/ c
O.R. = _______ = AxD÷ BxC
b/ d
=1 0r > 1 or < 1

OR=(a × d)/(b × c)
O.R=1 means that the odds of exposure among
the cases equal the odds of exposure among the
control, so no evidence of association between
exposure and outcome.
O.R >1 means that the cases have a higher odds
of exposure to the risk more than the control.
Ex.,People who ate food from the bakery had 3 times
the risk of developing hepatitis A compared to
people who did not eat food from the bakery

It is used as a descriptive statistic .

Fisher’s Exact Probability test and
the Chi-Square tests.

1- Study of obesity as a risk factor for DM.
2- The effect of high salt intake on Hypertensive
adults
3-The study of atrial fibrillation in middle aged
men during exercise.

Cohort means a group of people having a common
characteristic, e.g.
Example:
A group of female pharmacists working together in a
lab and some of them work on certain
chemicals(exposed) and others are not (unexposed).
They are followed up for a period of time and
studying disease(s) occurrence or unfavorable
outcomes of pregnancy for example.
Ex., A study of cancers and anomalies after atomic reactions

An example of an epidemiological question
that can be answered by the use of a
cohort study is:
Does exposure to X (say, smoking)
associate with outcome Y (say, lung
cancer)?

The exposure has already occurred
A cohort of people- with and without the exposure-
are followed up over a period of time to see who
develops the condition(s) of interest
A cohort study helps in establishing causation
With
exposure
Without
exposure
A cohort
group
followed

Exposure( start of the study) outcome

Methodology
A population of exposed and non-exposed
determined & followed prospectively for a certain
period of time “e.g. one year”
Incidence rate of an event for both exposed & non-
exposed calculated & relative risk is measured
Statistical analysis (by a test of significance )to
assess that the observed difference is real not due to
chance

TotalNot diseasedDiseased
A+bBAExposed
C+dDC ÷Not
exposed
B+dA+c

-RR= A/(A+B) / C/(C+D) it is a risk
measurement ,
It measures the association between incidence of disease and
certain exposure.
It answer the question:
How many times an exposed person is at risk of developing disease
compared to non- exposed?
RR= incidence among exposed / incidence
among non-exposed.

RR=1 : no association between exposure and the
disease. Incidence rate is the same in both
groups
RR = › 1 : positive association or increased risk ;
exposed group has higher incidence than non
exposed.
RR = ‹ 1: negative association or protective
effect; non exposed group has higher incidence.

AR= incidence among exposed – incidence
among non exposed.x100%
It study the rate of occurrence of the disease in
exposed individuals that can be attributed to the
exposure.(a real exposure and not by a chance)

Strengths Limitation
Valuable in rare
exposure
Less bias in recall &
observation
Allows for calculation
of Incidence rate” &
Relative risk
Inefficient in evaluating
rare diseases.
Costly & time
consuming
Loss of follow up may
affect validity of results.

Historical cohort
study:
(Better than case control
(incidence rate can be
calculated besides causal
association)
Ex. Cancer in certain
occupations
Nested case control
study:

Example
To determine if H. pylori (HP) infection was associated with the
development of gastric cancer,
Parsonnet et al (N Engl J Med. 1991 Oct 17;325(16):1127-31) first
identified a cohort of 128,992 persons who had been followed since
the mid-1960â€™s. Of the original cohort, 189 patients developed
gastric cancer. The investigators carried out a nested case-control
study by selecting all of the 189 gastric cancer patients as cases
and another 189 cancer-free individuals from the same cohort as
controls. HP infection status was determined using serum obtained
in the beginning of the follow-up. All total of 84% of the confirmed
gastric cancer cases had been infected previously with HP, while
only 61% of the controls had been infected; this indicated a positive
association between HP infection and gastric cancer risk.

In cross-sectional study both the risk factor and outcome
are present at time of study:
For example, a study finding an association between low
CD4 counts and HIV infection does not demonstrate
whether HIV infection lowers CD4 levels or low CD4 levels
predispose to HIV infection.

For example cases of DM (outcome at
present) are matched with a control group
and the
History of risk factors are traced
retrospectively (Exposure)

An experiment is a
A procedure carried out with the goal of
verifying, or establishing the validity of a
hypothesis. Experiments provide insight into
cause-and-effect.

The investigator can control both the population
& factors exposed
The experimenter control the independent
variable (ex. Treatment)
Then he measure the dependent variable (ex.,
health status).
It is the best epidemiological study design to
prove causation.

A true experiment is characterized by:
 It Randomly allocates the subjects bias.
 Controling all Confounding factors.
 A comparison group :
 Assessing the efficacy and safety of the intervention before
being a standard practice through a Comparison with a
standard intervention or therapy.

According to randomization; experimental studies are
divided into:
Experimental studies
•(Randomized studies
Quasi experimental studies
•Non randomized studies

A- Before and after (Pretest-post test) experimental
studies (with no control group):
B- Randomized controled trial (RCT) ;also called
(Pretest-posttest control group design)
Post-test only control group design) (rarely)
C- Clinical trials

A-Before and after (Pretest-post test) experimental
studies (with no control group):
The study subjects are randomly allocated
The researcher take the pretest measurements or
variables
Then intervene with the drug or vaccine and follow up
The researcher take the posttest measurements or
variables and then analyze the results

B- Randomized controlled trial (RCT) ;also called
(Pretest-posttest control group design)
N.B In a few cases, a pre-test is impossible
because the participants have already been exposed to
the treatment, or it would be too expensive or too
time-consuming.
And it is called ( post-test only control group design)

1. Identify the target population. ( inclusion &
exclusion criteria)
2. Consent (all participants)
3. Random Allocation
Each subject has an equal chance of being
included in the control or experimental group.

1. The experimental group Intervention or
drug
2. Control group Placebo or traditional
therapy
3. Follow up for a specified period
4. The outcome (preventive or curative) is
measured (often before and after intervention
5. The outcome is compared (experiment & control)

Examples of the outcomes in clinical
trials:
Systolic blood pressure in a
hypertension trial
Caesarean section rate in an obstetric
trial
Survival time in a cancer trial

Clinical Trials
Research activities in which:
Administration of therapeutic or
preventive intervention to evaluate its
safety and efficacy.

Clinical trials are useful in evaluating new:
Drugs in treatment of diseases.
Medical technology.
Methods of prevention.
Program for screening and diagnosis.

A-Pre-clinical trials in animals and lab (in vitro)
B-Testing in human through 4 phases”
Pharmacokinetics includes the study of the mechanisms
of absorption and distribution of an administered drug,
the chemical changes of the substance in the body (e.g.
by metabolic enzymes such as CYP or UGT enzymes), and
the effects and routes of excretion of the metabolites of
the drug.[1] Pharmacokinetics is often studied in
conjunction with pharmacodynamics , the study of a
drug's pharmacological effect on the body.

Volunteers Small
doses
Short
period
Close
supervision
Small doses taken for short
period
on severely ill or healthy
volunteers

Patient
Volunteers
Small
doses
longer
period
Close
supervision
sometimes this phase is
associated with a control group

Patient’s
consent
Usual
doses
longer
period
Control Gp
A classical phase,
Mostly this phase is associated
with randomization, double blind and a control
group.

Patient’s
consent
Usual
doses
long
period for
20 years
Control gp

1. Single blind : studied persons don’t know
whether they belong to the treatment or
control group
2. Double blind : studied persons &
experimenter not known
3. Triple blind : all the three studied
persons, experiment & statistician not
known

Single blind
Double blind
Experimental or
control group
Triple blind

1.(S )Allows for control of other variables
which affect outcomes under Investigation.
2.It is the best study design to prove
causation and prove validity
3.(L)
Insufficient number of subjects lead to
failure to detect true differences

Design a community trial
1. Single community design (before and after intervention).
2. Two Community design, one intervention community
and control.
3. One-to-many, the intervention community has several
control community.
4. In a many-to-many design there are study with multiple
intervention communities and multiple control
communities.

S It is useful in natural field circumstances.
L
Selection bias
People in the control community may receive the
intervention under study on their own because
There is no strict control as occurs in laboratory
experimental or animal studies is not possible with
humans.

The Salk vaccine trial carried out in 1954 had
200,000 experimental group and 200,000
control group
The aspirin-myocardial infarction study (as
therapeutic)
The intervention group received 1.0 gram of
aspirin daily whereas the reference group
received a placebo.

Two major types of quasi- experimental designs:
1. Pre test – post test, Nonequivalent control
group designs with-no random assignment to
groups
2. Before and after design- has pretest and
posttest with no comparison group with non
randomization to the study group

‫الرنة‬ ‫حيوانات‬ ‫برعاية‬ ‫القائمين‬

Definition:
It is a : Statistical technique that combine the
ﬁndings of a set of studies addressing
common research hypotheses to derive
definite conclusion
It Combines contradictory results of several
clinical studies on the same subject to derive
definite conclusion.

Study question: used to
Describe the best clinical approach to a problem.
Identify new directions for research.

Qualitative aspects
Developing research question to be
analyzed
Reviewing literature for relevant clinical
studies
Selecting variable common to the studies
Evaluating studies to identify similarities or
explain differences
Quantitative aspects : statistical analysis of results

Common designObjective
Cross sectionalPrevalence
CohortIncidence
Controlled trial, Cohort, case-
control, cross sectional
Cause (in order of
reliability)
CohortPrognosis (future or recovery
Controlled trialTreatment effect

Protection of human rights
Any person who is requested to consent to
participate as a subject in a research
study has the right to be
Informed of nature & purpose of
experiment
Given description of any risks to be
expected from experiment

Given explanation of expected benefits
from the experiment
Informed of medical treatment to subject
after experiment if complicated
Given an opportunity to ask any questions
concerning experiment or procedures
involved

Informed that consent may be withdrawn at
any time without prejudice.
Given a copy of a signed & dated written
consent form when required.
Given an opportunity to decide to consent
or not to consent to a medical experiment
without intervention

Statistics and biostatistics

More Related Content

What's hot (20)

Similar to Statistics and biostatistics (20)

Recently uploaded (20)

Statistics and biostatistics